On One Method of Constructing a Decision Rule in Pattern Recognition

The general problem of pattern recognition is to establish the theory and principles of building systems that share complex situations and events into classes. In this case, usually the number of possible situations in the classes is great and they are almost impossible to be remembered. The method of estimates' calculating contributes to a variety of scoring functions' building, which in turn, allows us to formulate and solve a wide class of multivariate data processing. In recent years more and more widely distributed methods of pattern recognition, in which the primary step is to calculate the balance of information and consideration of evidence. Based on the latest evidence being implemented, all further procedures until the distinction of classes or assign any object to one of the classes.

Перевод пока недоступен